In the distant past, issues of a journal would appear every two or three months, and publication date was no more precisely defined than this; you would only know that your paper had appeared when you received the journal (if you had subscribed) or your reprints (if you have ordered any).
Now each paper has a date of its on-line publication. These dates are not a lot of use for establishing priority, since the arXiv already does this, but are necessary for filling in data on institutional repositories.
This week, I had two papers published on the same day, something I don’t think has happened to me before. One, with R. A. Bailey and Tomas Nilson, was in the Australasian Journal of Combinatorics, where publication means that it comes with volume, issue and page numbers (this is a diamond open access journal which only publishes on-line). The other, with Bertalan Bodor and Csaba Szabó, was in Algebra Universalis, and is now “first on-line” for an indefinite period.
I have discussed them here and here.
By coincidence, both are in memorial volumes, for Anne Street and Tamás Schmidt respectively.
Lots of coincidences: you mentioned a similar phenomenon a while back (https://cameroncounts.wordpress.com/2015/04/02/two-publications/) — here both publications were also in special issues and one was also in the Australasian Journal of Combinatorics.
Correction: one of the previous examples wasn’t a special issue — it was your paper which was dedicated to Ákos Seress.
I’d completely forgotten that 😉